A systematic comparison of methodologies for the estimation of the serial interval

Area of Science:

• Epidemiology
• Infectious Disease Dynamics
• Public Health Surveillance

Background:

• Serial interval (SI) is key for understanding epidemic transmission and guiding interventions.
• Evaluating SI estimation methods is challenging due to the lack of real SI values.

Purpose of the Study:

• To develop a data-driven model simulating realistic epidemics to obtain true SI values for method comparison.
• To systematically compare the effectiveness of different SI estimation methods.

Main Methods:

• Developed a data-driven transmission model to simulate micro-transmission chains and obtain real SI values.
• Compared ML (contact tracing), BY (household imputation), and TM (aggregated data) methods.
• Analyzed data from different epidemic phases (early, peak, decreasing) and simulated noise.

Main Results:

• Household infector-infectee pairs accurately represent population SI.
• Early and peak phase household data provide precise SI estimations; decreasing phase data leads to underestimation.
• ML and BY methods outperform TM for general SI estimation, while TM is precise for mean SI during peak phases.
• Approximately 100 households are sufficient for accurate SI estimation.

Conclusions:

• Household data from early/peak epidemic phases are crucial for reliable SI estimation.
• The choice of SI estimation method (ML, BY, TM) and data collection phase impacts accuracy.
• Findings provide evidence for public health authorities to select appropriate SI estimation methods for policy-making.